CLINICAL ARTICLE

J Neurosurg 127:16–22, 2017

Blunt cerebrovascular injuries in severe traumatic brain

injury: incidence, risk factors, and evolution
Pierre Esnault, MD,1 Mickaël Cardinale, MD,1 Henry Boret, MD,1 Erwan D’Aranda, MD,1
Ambroise Montcriol, MD,1 Julien Bordes, MD,1 Bertrand Prunet, MD, PhD,1
Christophe Joubert, MD,2 Arnaud Dagain, MD,2,3 Philippe Goutorbe, MD,1 Eric Kaiser, MD,1,3 and
Eric Meaudre, MD1,3
1
Intensive Care Unit and 2Department of Neurosurgery, Sainte Anne Military Hospital, Toulon; and 3French Military Health Service
Academy Unit, École du Val-de-Grâce, Paris, France

OBJECTIVE  Blunt cerebrovascular injuries (BCVIs) affect approximately 1% of patients with blunt trauma. An antithrom-
botic or anticoagulation therapy is recommended to prevent the occurrence or recurrence of neurovascular events. This
treatment has to be carefully considered after severe traumatic brain injury (TBI), due to the risk of intracranial hemor-
rhage expansion. Thus, the physician in charge of the patient is confronted with a hemorrhagic and ischemic risk. The
main objective of this study was to determine the incidence of BCVI after severe TBI.
METHODS  The authors conducted a prospective, observational, single-center study including all patients with severe
TBI admitted in the trauma center. Diagnosis of BCVI was performed using a 64-channel multidetector CT. Charac-
teristics of the patients, CT scan results, and outcomes were collected. A multivariate logistic regression model was
developed to determine the risk factors of BCVI. Patients in whom BCVI was diagnosed were treated with systemic anti-
coagulation.
RESULTS  In total, 228 patients with severe TBI who were treated over a period of 7 years were included. The incidence
of BCVI was 9.2%. The main risk factors were as follows: motorcycle crash (OR 8.2, 95% CI 1.9–34.8), fracture involving
the carotid canal (OR 11.7, 95% CI 1.7–80.9), cervical spine injury (OR 13.5, 95% CI 3.1–59.4), thoracic trauma (OR 7.3,
95% CI 1.1–51.2), and hepatic lesion (OR 13.3, 95% CI 2.1–84.5). Among survivors, 82% of patients with BCVI received
systemic anticoagulation therapy, beginning at a median of Day 1.5. The overall stroke rate was 19%. One patient had an
intracranial hemorrhagic complication.
CONCLUSIONS  Blunt cerebrovascular injuries are frequent after severe TBI (incidence 9.2%). The main risk factors are
high-velocity lesions and injuries near cervical arteries.
https://thejns.org/doi/abs/10.3171/2016.4.JNS152600
KEY WORDS  blunt cerebrovascular injury; traumatic brain injury; cervical artery dissection; anticoagulation therapy

B
lunt cerebrovascular injuries (BCVIs) are caused
by penetration of circulating blood into the arterial
wall. Intramural hematoma creates a false lumen
that leads to stenosis, occlusion, or pseudoaneurysm of the
affected artery.2 In the past decade BCVIs have been in-
creasingly recognized due to initiation of different screen-
ing protocols, with an incidence of approximately 1% in
patients with blunt trauma.15,24,25 These lesions are rare but
potentially devastating because they can lead to neurovas-
cular events.5,6
Mechanisms of these events are dual: usually embolic,
from intraluminal thrombi forming at the site of the inti-
mal tear, or hemodynamic, by arterial stenosis or vessel
occlusion that may lead to low-flow infarcts.21 It is now
recognized that a latent, asymptomatic period between in-
jury and development of stroke frequently exists.17,20 An
antithrombotic or anticoagulation therapy established dur-
ing this period may improve neurological outcome and
mortality rates.8,19,23 However, systemic anticoagulation
carries a significant risk of hemorrhage in patients who
suffer a trauma. Especially after traumatic brain injury
(TBI), anticoagulation should not be given within 3 days
of injury for moderate- or high-risk intracranial hemor-
rhage (ICH), because the administration of heparin, even
at a prophylactic dose, may cause ICH expansion.1
Thus, in case of BCVI associated with TBI, the physi-

ABBREVIATIONS  BCVI = blunt cerebrovascular injury; CAI = carotid artery injury; CTA = CT angiography; DSA = digital subtraction angiography; GCS = Glasgow Coma
Scale; GOS = Glasgow Outcome Scale; ICH = intracranial hemorrhage; ICP = intracranial pressure; IQR = interquartile range; ISS = Injury Severity Score; LMWH = low-
molecular-weight heparin; ROC = receiver operating characteristic; TBI = traumatic brain injury; VAI = vertebral artery injury.
SUBMITTED  November 12, 2015.  ACCEPTED  April 29, 2016.
INCLUDE WHEN CITING  Published online July 29, 2016; DOI: 10.3171/2016.4.JNS152600.

16 J Neurosurg  Volume 127 • July 2017 ©AANS, 2017


cian treating the patient is challenged by a hemorrhagic
and an ischemic risk, both with potentially devastating
neurological consequences. However, this association is
poorly studied in the medical literature. The only avail-
able data showed a BCVI incidence of 3.9% in a subgroup
of head-injured patients (isolated cranial fractures with or
without intracranial hematoma).24 The main objective of
this study was to describe the incidence and characteris-
tics of BCVI in patients with severe TBI.
Methods
Study Design and Patient Selection
We conducted a 7-year prospective, observational,
single-center study at the Sainte Anne Military Hospital
of Toulon, France. Between January 2007 and December
2013, all intubated patients with TBI admitted to the ICU
were eligible. Inclusion criteria were as follows: patients
older than 18 years who had severe TBI, defined as a mini-
mum Glasgow Coma Scale (GCS) score of ≤ 8 after re-
suscitation. Patients suffering from penetrating trauma and
those in whom no CT scan was obtained within the first 24
hours were excluded from the study.
All patients were sedated, intubated, and received me-
chanical ventilation in accordance with the international
guidelines and with the following objectives: head eleva-
tion 30°, PaO2 > 85 mm Hg, PaCO2 between 35 and 45 mm
Hg, natremia between 140 and 145 mEq/L, targeted tem-
perature < 37.5°C, intracranial pressure (ICP) < 20 mm
Hg, and cerebral perfusion pressure > 60 mm Hg. If ICP
increased, the protocol included moderate hypocapnia
(PaCO2 between 30 and 35 mm Hg), increase of sedation
with propofol, muscle paralysis with cisatracurium, and
moderate hypothermia (34°C–35°C). Third-line therapy
included barbiturate infusion and/or unilateral craniecto-
my.9–14
The institutional review board approved the study and
waived the requirement for informed consent from the pa-
tient or patient’s next of kin, given the observational nature
of the study.
Data Collected and Diagnostic Modalities
Patient characteristics, mechanisms of injury, initial
GCS score, results of whole-body CT scan, location and
grade of BCVI, timing and type of initial therapy, follow-
up imagery, evolution of the disease, in-hospital mortality,
and outcome according to the Glasgow Outcome Scale
(GOS) at 6 months and 1 year were collected. A thoracic
trauma included all chest lesions except for thoracic ver-
tebral fractures.
Primary screening of BCVI was done using a 64-chan-
nel CT scanner with CT angiography (CTA) (General
Electric Medical Systems). Patients were positioned with
their arms down to the side and underwent head and cer-
vical spine studies without contrast. Scout images were
obtained and the carotid vessels identified. Radiopaque
iodinated contrast dye injection at 4 ml/sec was given
(Omnipaque 65 ml). When a contrast blush was noted at
the origin of the carotid arteries, the CTA head/neck scan
was initiated from the clavicles and continued to the apex
of the calvaria. Images were acquired at 0.625-mm-thick
Blunt cerebrovascular injuries in severe traumatic brain injury
slices and 0.625-mm intervals. If necessary, patient whole-
body imaging was completed. Digital reconstruction was
performed immediately after the completion of CT scans.
Sagittal and coronal images were reconstructed at 2-mm
by 2-mm intervals, whereas axial images were at 1-mm by
1-mm intervals.
In the situations indicated by the institutional protocol,
digital subtraction angiography (DSA) (Philips Health-
care) was performed by an experienced neuroradiologist.
Selective single-plane DSA of the aortic arch and each
of the subclavian arteries, and biplanar 4-vessel DSA of
the cerebrovascular arteries were performed. The attend-
ing neuroradiologist’s interpretations of CTA and/or DSA
were used to determine BCVI location and grade accord-
ing to the Biffl et al. grading scale.7 Grade I injuries were
defined as irregularity of the vessel wall or a dissection
≤ 25% luminal stenosis. Grade II injuries included an in-
traluminal thrombus, intimal flap, or dissections or intra-
mural hematomas with > 25% luminal stenosis. Grade III
injuries were pseudoaneurysms. Grade IV injuries were
vessel occlusions. Grade V injuries were complete vessel
transection with free contrast extravasation or arteriove-
nous fistulas.
Treatment Methods
Patients in whom BCVI was diagnosed were treated
with systemic anticoagulation (unfractionated heparin or
low-molecular-weight heparin [LMWH]). The beginning
point and type of systemic anticoagulation were chosen at
the discretion of the treating physician, while respecting
the benefit-risk ratio. In the case of unfractionated hepa-
rin, partial thromboplastin time was maintained at 45–60
seconds. In case of LMWH, anti–factor Xa activity was
maintained at 0.4–0.6 IU/ml. A small number of patients
received no systemic treatment due to endovascular treat-
ment, major bleeding risk, or withdrawal of care.
Study End Points
The primary end point of this study was to determine
the incidence of BCVI after severe TBI. Secondary end
points were to determine risk factors of BCVI in this
population and to describe the evolution of the disease in
patients with BCVI.
Statistical Analysis
Statistical analysis was performed with XLSTAT ver-
sion 2015.3.01 (Addinsoft). Continuous data were reported
as the mean ± SD, or the median with the interquartile
range (IQR) (25th–75th percentile) when data were not
normally distributed. Nominal variables are reported as
numbers and proportions (%).
For this observational study with consecutive sampling,
a = 0.05, b = 0.2, and an estimated incidence of 4% for
BCVI after severe TBI, we planned to include at least 191
patients to yield an expected incidence rate of 5%.24
A univariate analysis was conducted using the chi-
square test or Fisher’s exact test to compare categorical
variables, and the Mann-Whitney test or Student t-test to
compare groups for continuous variables (for comparison
of medians and comparison of means, respectively). In-
J Neurosurg  Volume 127 • July 2017 17
P. Esnault et al.

FIG. 1. Pie chart showing the mechanism of injury of enrolled patients.

dependent factors associated with BCVI were identified with vertebral artery injury (VAI), and 1 (5%) with injury
using a logistic regression model. All parameters with p to both of them. The incidence of BCVI was 9.2% of all
value < 0.05 on univariate analysis and clinical relevance admissions for severe TBI. A CAI occurred in 19 vessels
were included in the multivariate regression model. The (7 left, 6 right, and 3 with bilateral lesions). A VAI oc-
Hosmer-Lemeshow goodness-of-fit test and the area under curred in 6 arteries (2 left, 4 right). The majority of inju-
the receiver operating characteristic (ROC) curve were ries were dissections (Grade I or II injuries): 17 vessels
used to evaluate the overall fit of the final model. The final (68%). Table 1 provides the location and grade of detected
model expressed the odds ratio and 95% confidence in- injuries.
tervals. For all tests, p < 0.05 was considered statistically Comparisons of patients with and without BCVI are
significant. displayed in Table 2. Notably, patients with BCVI were
more severely injured, with a median ISS of 41 (IQR 29–
Results 50) versus 22 (IQR 16–32) (p < 0.0001) for those without
BCVI. Patients with BCVI also had initial hemodynamic
Patient Population shock and/or hypoxia (SpO2 < 90% during > 5 minutes)
During the study period, 243 intubated patients with more frequently: 12 of 21 (57%) versus 44 of 207 (21%)
TBI were admitted to the ICU. Fifteen patients were ex- (p = 0.0002), and 9 (43%) versus 29 (14%) (p = 0.001),
cluded: 10 due to mild or moderate TBI (minimum GCS respectively. The number and type of intracranial lesions
score of > 8 after resuscitation); 4 due to penetrating trau- were comparable between the groups. However, associat-
ma; and 1 due to the absence of a CTA study obtained
within the first 24 hours. The remaining 228 were includ-
ed. The most common mechanism of injury was motor- TABLE 1. Grade and location of BCVIs
cycle crash (30.1%) (Fig. 1). The median age of the popu-
lation was 37 years (IQR 23–55 years), the median Injury Grade Grade Grade Grade Grade Total
Severity Score (ISS) was 24 (IQR 16–34), and the median Vessel I II III IV V (%)
initial GCS score was 6 (IQR 4–8). There were 179 male Rt CAI 2 4 2 0 1 9 (36)
(78.5%) and 49 female (21.5%) patients. Lt CAI 2 5 0 1 2 10 (40)
Rt VAI 1 1 0 2 0 4 (16)
Blunt Cerebrovascular Injuries
Lt VAI 2 0 0 0 0 2 (8)
A BCVI was detected in 21 patients. These included 15
patients (71%) with carotid artery injury (CAI), 5 (24%) Total (%) 7 (28) 10 (40) 2 (8) 3 (12) 3 (12) 25 (100)

18 J Neurosurg  Volume 127 • July 2017

 Blunt cerebrovascular injuries in severe traumatic brain injury

TABLE 2. Characteristics of patients according to the presence TABLE 3. Multivariate logistic regression model for risk factors
of BCVI* of BCVI*
Patients w/o Patients w/ Variable OR 95% CI p Value
Variable BCVI, n = 207 BCVI, n = 21 p Value
ISS 1.04 1.01–1.08 0.02
Age in yrs 38 [23–58] 31 [21–46] 0.18 Initial GCS score 0.82 0.63–1.07 0.14
Male sex 159 (77) 20 (95) 0.05 Motorcycle crash 8.17 1.92–34.78 0.004
SAPS II 42 [33–53] 45 [36–61] 0.36 Basilar skull fracture 2.18 0.46–10.27 0.32
ISS 22 [16–32] 41 [29–50] <0.0001 Fracture involving the 11.65 1.68–80.86 0.01
Brain injury carotid canal
 EDH 35 (17) 2 (10) 0.38 Cervical spine injury 13.55 3.09–59.36 0.001
 SDH 110 (53) 8 (38) 0.19 Thoracic lesion 7.35 1.06–51.19 0.04
 IPH 28 (14) 2 (10) 0.61 Hepatic lesion 13.26 2.08–84.54 0.006
 SAH 99 (48) 10 (48) 0.99 *  Area under the ROC curve: 0.94; Hosmer-Lemeshow test: χ2 = 3.3, df = 8,
 Contusion 60 (29) 2 (10) 0.06 p = 0.92.
 Petechial 85 (41) 11 (52) 0.32
 IVH 29 (14) 2 (10) 0.57 ed lesions were more frequent in patients with BCVI (90%
 Shift 60 (29) 3 (14) 0.15 vs 66%, p = 0.02). Patients with BCVI had more thoracic
  Brain swelling 60 (29) 4 (19) 0.33 trauma (90% vs 45%, p < 0.0001), cervical spine injuries
  Marshall category 2 [2–5] 2 [2–3] 0.11
(43% vs 8%, p < 0.0001), and hepatic lesions (24% vs 3%,
p < 0.0001). Initial GCS scores were lower in patients with
Associated lesions 136 (66) 19 (90) 0.02 BCVI: median 4 (IQR 3–6) versus 7 (IQR 4–9) (p = 0.005).
Facial trauma 91 (44) 12 (57) 0.25 The predominant mechanism of injury in the BCVI group
  Upper lesions 34 (16) 3 (14) 0.8 was motorcycle crash (57% vs 28%, p = 0.005).
  Middle lesions 63 (30) 8 (38) 0.47
  Lower lesions 14 (7) 1 (5) 0.72 Risk Factors for BCVI
  Basilar skull fracture 39 (19) 9 (43) 0.01 After multivariate logistic regression analysis, major
  Involving carotid canal 12 (6) 6 (29) 0.0002 significant predictors of BCVI were as follows: motorcy-
cle crash (OR 8.2; 95% CI 1.9–34.8, p = 0.004); fracture
Spine injury 38 (18) 11 (52) 0.0002
involving the carotid canal (OR 11.7; 95% CI 1.7–80.9, p
  Cervical spine 16 (8) 9 (43) <0.0001 = 0.01); cervical spine injury (OR 13.5; 95% CI 3.1–59.4,
  Thoracic spine 17 (8) 4 (19) 0.1 p = 0.001); thoracic trauma (OR 7.3; 95% CI 1.1–51.2, p =
  Lumbar spine 17 (8) 3 (14) 0.35 0.04); and hepatic lesion (OR 13.3; 95% CI 2.1–84.5, p =
Thoracic injury 94 (45) 19 (90) <0.0001 0.006). Other variables included in the model are shown
  Pulmonary contusions 72 (35) 16 (76) 0.0002 in Table 3. The accuracy of the model was excellent, with
 Hemothorax 23 (11) 3 (14) 0.66
an area under the ROC curve of 0.94. The Hosmer-Lem-
eshow test demonstrated a good model fit (c2 = 3.3, df =
 Pneumothorax 43 (21) 5 (24) 0.75 8, p = 0.92).
  Rib fractures 48 (23) 10 (48) 0.01
Abdominal trauma 31 (15) 6 (29) 0.11 Treatment, Stroke Rate, and Hemorrhagic Complications
  Hepatic lesion 7 (3) 5 (24) <0.0001 Four of the 21 patients with BCVI died prematurely be-
  Splanchnic lesion 10 (5) 1 (5) 0.99 fore definitive treatment; these individuals had significant
Pelvic trauma 32 (15) 4 (19) 0.67 associated injuries. Among the 17 survivors, 14 patients
Orthopedic lesions 71 (34) 12 (57) 0.04 (82%) had medical treatment with systemic anticoagula-
tion therapy. Four patients received unfractionated heparin
Neurological symptoms
and 10 had LMWH. The beginning of systemic anticoag-
  Minimum GCS score 6 [4–8] 4 [3–6] 0.04 ulation was at a median of Day 1.5 (IQR Day 1–4). Seven
  Lateralizing deficit 114 (55) 8 (38) 0.14 patients (41%) were treated with systemic anticoagulation
 Anisocoria 74 (36) 4 (19) 0.12 within the first 24 hours after admission. Two patients
  Bilat mydriasis 20 (10) 3 (14) 0.5 (12%) had endovascular interventions with coils (for treat-
Initial hemoglobin in g/dl 12.7 [11.2–14.2] 12.3 [8.9–14.2] 0.5 ment of Grade III right CAI and Grade V bilateral CAI,
respectively). One patient (6%) had just chemoprophylaxis
EDH = epidural hematoma; IPH = intraparenchymal hematoma; IVH = intraven- for deep vein thrombosis started at Day 5 (Grade I right
tricular hemorrhage; SAH = subarachnoid hemorrhage; SAPS II = Simplified VAI with severe ICH). Four patients with BCVI suffered
Acute Physiology Score; SDH = subdural hematoma.
*  Data are presented as the number (%) or the median [IQR: 25th–75th
an ischemic stroke, for an overall stroke rate of 19%. One
percentile]. stroke was present on arrival (Grade IV left CAI), 1 de-
veloped before introduction of systemic anticoagulation
therapy (Grade II right CAI), and 2 developed despite
early anticoagulation therapy (2 patients with Grade II left

J Neurosurg  Volume 127 • July 2017 19

P. Esnault et al.

CAI). One patient (6%) had an ICH complication due to TABLE 4. Clinical follow-up and outcomes according to the
systemic anticoagulation. This patient developed a surgi- presence of BCVI*
cally evacuated epidural hematoma at Day 5 following the Patients w/o Patients w/ p
insertion of an ICP monitoring device. Variable BCVI BCVI Value

Posttreatment Outcomes Barbiturate therapy 51 (25) 2 (10) 0.12

Clinical follow-up was not different between patients Therapeutic hypothermia 67 (32) 5 (24) 0.42
with and those without BCVI, in terms of intracranial Decompressive craniectomy 44 (21) 4 (19) 0.81
hypertension episodes, use of barbiturate or therapeutic Intracranial hypertension 91 (44) 13 (62) 0.12
hypothermia, or number of decompressive craniectomies frequency
(Table 4). The mortality rate was comparable between the Day of sedation interruption 3 [1–8] 4 [3–10] 0.4
groups: 33% in patients with versus 26% in patients with- Duration of MV in days 7 [2–13] 9 [3–16] 0.38
out BCVI (p = 0.48). The incidence of poor outcome (de- Tracheostomy 75 (36) 8 (38) 0.87
fined as a GOS score between 1 and 3) was not different
at 6 months and 1 year between patients with or without Early-onset pneumonia 89 (43) 9 (43) 0.99
BCVI: 50% versus 45% (p = 0.66) and 47% versus 43% (p ICU length of stay in days 9 [3–17] 11 [3–20] 0.69
= 0.7), respectively. Hospital length of stay in days 17 [7–31] 22 [5–39] 0.77
In-hospital mortality 54 (26) 7 (33) 0.48
Discussion GOS Score 1–3 at 6 mos† 92/205 (45) 10/20 (50) 0.66
GOS Score 1–3 at 1 yr† 80/187 (43) 9/19 (47) 0.7
To our knowledge, this article is the first to focus on
BCVI in patients with severe TBI. Our results highlight a MV = mechanical ventilation.
rate of 9.2% for BCVI in this particular population. BCVI *  Data are presented as the number (%) or the median [IQR: 25th–75th
is increasingly recognized, due to advancements in screen- percentile].
ing and imaging technology. In the recent literature, the †  Some patients were lost to follow-up (often foreign patients on holiday in
our region): 3 patients at 6 months (1 in the BCVI group and 2 in the non-BCVI
incidence of BCVI in all patients with blunt trauma has group); and 22 patients at 1 year (2 in the BCVI group and 20 in the non-BCVI
increased from 0.1% to 1.7%.4,8,15,24,25,27,31,35 It is now recog- group).
nized that patients with facial fractures, cervical spine inju-
ry, or thoracic lesions have an increased risk of BCVI.4,25,32
Our study strongly suggests that patients with severe skull fracture and initial GCS score.4,27 This is probably
TBI are a very high-risk population for BCVI. This is due to the particular population of the study. Indeed, only
probably due to common physiological mechanisms. In- patients with severe TBI (GCS score of ≤ 8) were included.
deed, the main mechanisms of BCVI are the following: This limits the power of the study to recognize GCS score
hyperextension with contralateral rotation of the head, as an independent risk factor. Concerning basilar skull
laceration of the artery by adjacent fractures, direct blow fractures, they are too frequent in our population (21%) to
to the neck, and direct intraoral trauma with a hard ob- find a significant difference in multivariate analysis.
ject.7,16,17,26,34 These mechanisms generally result from a Untreated vascular injuries could have an overall stroke
high-speed deceleration crash with cervical hyperexten- rate approaching 40%–60%, with a stroke-related mor-
sion or hyperflexion, or direct blunt trauma to the neck, tality as high as 50%.4,18,22,35 It is now well established
face, head, or upper chest. For TBI, similar mechanisms of that medical therapy (with anticoagulation or antiplate-
injury exist: impact—in which the immediate force makes let agents) that is started before the onset of neurological
contact with the skull and leads to injury; and impulse— symptoms could reduce the occurrence of neurological
in which a force causes head movement without directly events by a factor of 6.28,35 The stroke rate could be as low
acting upon the head.3 Moreover, in our study, patients as 4% in treated patients.28,35 Consistent with other stud-
with severe TBI frequently had injuries to the upper body: ies, our overall stroke rate was 19%, despite the frequent
facial fractures (103/228; 45%), cervical spine injuries and early use of anticoagulation therapy. Some could
(25/228; 11%), and thoracic lesions (113/228; 50%). All of have doubts concerning the safety of this therapy in pa-
these can explain the very high incidence of BCVI found tients with severe TBI. Recently, Callcut et al. examined
in this population. the safety and efficacy of early pharmacological treatment
The main risk factors of BCVI after severe TBI are for patients with both BCVI and traumatic neurological
fracture involving the carotid canal, cervical spine injury, injury, including TBI and/or spinal cord injury.18 They
and thoracic trauma. This is consistent with the litera- found that early medical therapy significantly reduces the
ture.4,25,27,30 In fact, Franz et al. demonstrated that cervi- stroke rate (from 57% to 4%), without increasing the risk of
cal spine and thoracic injuries are significantly associated ICH expansion (from 5% to 6%). However, pharmacologi-
with BCVI in all blunt trauma admissions (OR 5.45 and cal treatment was only initiated at a median of Day 3, and
1.98, respectively).25 We also found new significant predic- their patient population was different, including individu-
tors of BCVI: ISS and, surprisingly, hepatic lesions. This als with mild or moderate TBI.
is certainly reflecting the violence of the crash. This hy- In our study, anticoagulation therapy was started soon-
pothesis is confirmed by the fact that a motorcycle crash is er, at Day 1.5, and 41% of patients were treated on arrival
also an independent risk factor for BCVI. However, 2 well- in the ICU. We chose to perform this therapy early be-
established predictors of BCVI were not retrieved: basilar cause the latent period between injury and ischemia was

20 J Neurosurg  Volume 127 • July 2017


observed mostly within 10–72 hours.8,29 Despite this very
early therapy, only 1 ICH was reported following an ICP
monitoring. There was no hemorrhagic deterioration of
initial lesions due to pharmacological exposure. The lit-
erature demonstrated no difference in the efficacy of anti-
platelet and anticoagulation therapies at preventing stroke
in patients suffering from BCVI.28 In our institution, sys-
temic anticoagulation is preferred for 2 reasons: 1) its ef-
fects can be easily monitored (by partial thromboplastin
time or anti–factor Xa activity); and 2) in case of hem-
orrhagic complication or emergency surgery, this therapy
can be quickly counteracted by the administration of prot-
amine sulfate.
Study Limitations
Our study had several limitations. First, the relatively
small number of patients with BCVI limits the power of
the multivariate analysis. Second, our results cannot be
transposed to all ICUs or trauma centers, due to the single-
center recruitment. Finally, CTA by 64-channel multide-
tector scanner was used for diagnosing or excluding BCVI.
A previous study using 32-channel CTA showed an inad-
equate sensitivity of 51% compared with DSA.22 However,
compared with the 32-channel CT scanner, Paulus et al.
recently highlighted that 64-channel CTA demonstrated a
significantly improved sensitivity of 68%.33 Specificity and
negative predictive value were 92% and 97%, respectively.
Despite better results with the 64-channel CTA, we could
have missed a certain number of patients with true BCVI
in our study population, and therefore the true incidence
may be higher. However, Paulus and colleagues found that
62% of the false-negative findings involved low-grade in-
jury (Grade I).33 These low-grade injuries probably have
a lower impact in patients and are rarely complicated by
neurovascular events. Indeed, Biffl et al. highlighted their
finding that two-thirds of mild intimal injuries (Grade I)
healed at Day 7, regardless of therapy.7 Thus, our results
seem to stay relevant in daily practice.
Conclusions
Despite these limitations, the present work is, to our
knowledge, the first designed to evaluate the incidence of
BCVI in patients with severe TBI. In this particular popula-
tion, the incidence of BCVI was very high, approximately
9%. The main risk factors reflected high-velocity injuries
like those sustained in motorcycle crashes or leading to
hepatic lesions; or were related to cervical artery proxim-
ity injuries such as fractures involving the carotid canal or
cervical spine injuries. In addition, our study demonstrates
that early administration of medical therapy could be safe.
However, prospective randomized multicenter studies are
mandatory to confirm this last point.
References
  1. Abdel-Aziz H, Dunham CM, Malik RJ, Hileman BM: Tim-
ing for deep vein thrombosis chemoprophylaxis in traumatic
brain injury: an evidence-based review. Crit Care 19:96,
2015
  2. Arthurs ZM, Starnes BW: Blunt carotid and vertebral artery
injuries. Injury 39:1232–1241, 2008
Blunt cerebrovascular injuries in severe traumatic brain injury
  3. Bauer D, Tung ML, Tsao JW: Mechanisms of traumatic brain
injury. Semin Neurol 35:e14–e22, 2015
  4. Berne JD, Cook A, Rowe SA, Norwood SH: A multivariate
logistic regression analysis of risk factors for blunt cerebro-
vascular injury. J Vasc Surg 51:57–64, 2010
  5. Berne JD, Norwood SH, McAuley CE, Vallina VL, Creath
RG, McLarty J: The high morbidity of blunt cerebrovascular
injury in an unscreened population: more evidence of the
need for mandatory screening protocols. J Am Coll Surg
192:314–321, 2001
  6. Biffl WL, Moore EE, Elliott JP, Ray C, Offner PJ, Franciose
RJ, et al: The devastating potential of blunt vertebral arterial
injuries. Ann Surg 231:672–681, 2000
  7. Biffl WL, Moore EE, Offner PJ, Brega KE, Franciose RJ,
Burch JM: Blunt carotid arterial injuries: implications of a
new grading scale. J Trauma 47:845–853, 1999
  8. Biffl WL, Moore EE, Ryu RK, Offner PJ, Novak Z, Coldwell
DM, et al: The unrecognized epidemic of blunt carotid arte-
rial injuries: early diagnosis improves neurologic outcome.
Ann Surg 228:462–470, 1998
  9. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the manage-
ment of severe traumatic brain injury. I. Blood pressure and
oxygenation. J Neurotrauma 24 (Suppl 1):S7–S13, 2007
(Erratum in J Neurotrauma 25:276–278, 2008)
10. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the management
of severe traumatic brain injury. II. Hyperosmolar therapy.
J Neurotrauma 24 (Suppl 1):S14–S20, 2007 (Erratum in J
Neurotrauma 25:276–278, 2008)
11. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the management
of severe traumatic brain injury. VIII. Intracranial pressure
thresholds. J Neurotrauma 24 (Suppl 1):S55–S58, 2007
(Erratum in J Neurotrauma 25:276–278, 2008)
12. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the manage-
ment of severe traumatic brain injury. IX. Cerebral perfusion
thresholds. J Neurotrauma 24 (Suppl 1):S59–S64, 2007
(Erratum in J Neurotrauma 25:276–278, 2008)
13. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the management
of severe traumatic brain injury. X. Brain oxygen monitoring
and thresholds. J Neurotrauma 24 (Suppl 1):S65–S70, 2007
(Erratum in J Neurotrauma 25:276–278, 2008)
14. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond
FF, Harris OA, Hartl R, et al: Guidelines for the management
of severe traumatic brain injury. XIV. Hyperventilation. J
Neurotrauma 24 (Suppl 1):S87–S90, 2007 (Erratum in J
Neurotrauma 25:276–278, 2008)
15. Bromberg WJ, Collier BC, Diebel LN, Dwyer KM, Holevar
MR, Jacobs DG, et al: Blunt cerebrovascular injury practice
management guidelines: the Eastern Association for the Sur-
gery of Trauma. J Trauma 68:471–477, 2010
16. Burlew CC, Biffl WL: Blunt cerebrovascular trauma. Curr
Opin Crit Care 16:587–595, 2010
17. Burlew CC, Biffl WL: Imaging for blunt carotid and verte-
bral artery injuries. Surg Clin North Am 91:217–231, 2011
18. Callcut RA, Hanseman DJ, Solan PD, Kadon KS, Ingalls NK,
Fortuna GR, et al: Early treatment of blunt cerebrovascular
injury with concomitant hemorrhagic neurologic injury is
safe and effective. J Trauma Acute Care Surg 72:338–346,
2012
19. Cothren CC, Biffl WL, Moore EE, Kashuk JL, Johnson JL:
Treatment for blunt cerebrovascular injuries: equivalence of
anticoagulation and antiplatelet agents. Arch Surg 144:685–
690, 2009
20. Cothren CC, Moore EE, Biffl WL, Ciesla DJ, Ray CE Jr,
Johnson JL, et al: Anticoagulation is the gold standard thera-
J Neurosurg  Volume 127 • July 2017 21
P. Esnault et al.
py for blunt carotid injuries to reduce stroke rate. Arch Surg
139:540–546, 2004
21. Debette S, Leys D: Cervical-artery dissections: predisposing
factors, diagnosis, and outcome. Lancet Neurol 8:668–678,
2009
22. DiCocco JM, Emmett KP, Fabian TC, Zarzaur BL, Williams
JS, Croce MA: Blunt cerebrovascular injury screening with
32-channel multidetector computed tomography: more slices
still don’t cut it. Ann Surg 253:444–450, 2011
23. Fabian TC, Patton JH Jr, Croce MA, Minard G, Kudsk KA,
Pritchard FE: Blunt carotid injury. Importance of early di-
agnosis and anticoagulant therapy. Ann Surg 223:513–525,
1996
24. Fleck SK, Langner S, Baldauf J, Kirsch M, Kohlmann T,
Schroeder HWS: Incidence of blunt craniocervical artery
injuries: use of whole-body computed tomography trauma
imaging with adapted computed tomography angiography.
Neurosurgery 69:615–624, 2011
25. Franz RW, Willette PA, Wood MJ, Wright ML, Hartman JF:
A systematic review and meta-analysis of diagnostic screen-
ing criteria for blunt cerebrovascular injuries. J Am Coll
Surg 214:313–327, 2012
26. Fusco MR, Harrigan MR: Cerebrovascular dissections: a
review. Part II: blunt cerebrovascular injury. Neurosurgery
68:517–530, 2011
27. Jacobson LE, Ziemba-Davis M, Herrera AJ: The limitations
of using risk factors to screen for blunt cerebrovascular inju-
ries: the harder you look, the more you find. World J Emerg
Surg 10:46, 2015
28. Markus HS, Hayter E, Levi C, Feldman A, Venables G, Nor-
ris J: Antiplatelet treatment compared with anticoagulation
treatment for cervical artery dissection (CADISS): a ran-
domised trial. Lancet Neurol 14:361–367, 2015
29. Mayberry JC, Brown CV, Mullins RJ, Velmahos GC: Blunt
carotid artery injury: the futility of aggressive screening and
diagnosis. Arch Surg 139:609–613, 2004
30. McKinney A, Ott F, Short J, McKinney Z, Truwit C: Angio-
graphic frequency of blunt cerebrovascular injury in patients
with carotid canal or vertebral foramen fractures on multide-
tector CT. Eur J Radiol 62:385–393, 2007
31. Miller PR, Fabian TC, Croce MA, Cagiannos C, Williams JS,
Vang M, et al: Prospective screening for blunt cerebrovascu-
22 J Neurosurg  Volume 127 • July 2017
lar injuries: analysis of diagnostic modalities and outcomes.
Ann Surg 236:386–395, 2002
32. Mundinger GS, Dorafshar AH, Gilson MM, Mithani SK,
Manson PN, Rodriguez ED: Blunt-mechanism facial frac-
ture patterns associated with internal carotid artery injuries:
recommendations for additional screening criteria based on
analysis of 4,398 patients. J Oral Maxillofac Surg 71:2092–
2100, 2013
33. Paulus EM, Fabian TC, Savage SA, Zarzaur BL, Botta V,
Dutton W, et al: Blunt cerebrovascular injury screening with
64-channel multidetector computed tomography: more slices
finally cut it. J Trauma Acute Care Surg 76:279–285, 2014
34. Sliker CW: Blunt cerebrovascular injuries: imaging with mul-
tidetector CT angiography. Radiographics 28:1689–1710,
2008
35. Stein DM, Boswell S, Sliker CW, Lui FY, Scalea TM: Blunt
cerebrovascular injuries: does treatment always matter? J
Trauma 66:132–144, 2009
Disclosures
The authors report no conflict of interest concerning the materi-
als or methods used in this study or the findings specified in this
paper.
Author Contributions
Conception and design: Esnault, Cardinale, Boret. Acquisition
of data: Esnault, Cardinale, Boret, D’Aranda, Montcriol, Bordes.
Analysis and interpretation of data: Esnault, Boret, Prunet,
Joubert. Drafting the article: Esnault, Boret, Dagain, Goutorbe,
Kaiser, Meaudre. Critically revising the article: Boret, Dagain,
Goutorbe, Kaiser, Meaudre. Reviewed submitted version of man-
uscript: Esnault. Approved the final version of the manuscript on
behalf of all authors: Esnault.
Correspondence
Pierre Esnault, Service de Réanimation—Brûlés, Hôpital
d’Instruction des Armées Sainte Anne, Blvd. Sainte Anne, Toulon
83000, France. email: pierre.esnault@gmail.com.